In the sprawling ecosystem of Bittensor, where dozens of subnets compete to define the future of decentralized intelligence, few have managed to balance engineering depth with real-world traction quite like Gopher, grounded on Bittensor’s Subnet 42.
The subnet, led by Brendan Playford, has quietly evolved from a data-scraping experiment into a fully-fledged infrastructure layer for verifiable AI data.
On episode 143 of Hash Rate, Mark Jeffrey hosts Playford to trace that transformation; from the project’s early days as Masa to its growing influence as Gopher.
A New Chapter for Subnet 42
“Let’s talk about Gopher,” Mark began. “I know this used to be called Masa. It feels like more than just a rebrand. What’s really changed?”
Brendan smiled, ready to explain. “We launched Subnet 42 over a year ago as a data scraping subnet under the Masa brand. At the time, our focus was gathering data from platforms like X, Reddit, and TikTok, but not for pre-training models. We wanted to give large language models context through RAG and tool configuration.”
That early focus evolved as the team realized that “data scraping” alone wasn’t enough to stand out in a crowded AI landscape. Around August, Brendan and his team decided to pivot. “We saw traction in one vertical that really made sense,” he said. “That’s when we rebranded as Gopher, a move that wasn’t just cosmetic. It was about clarity of story and value.”
What Makes Gopher Different
Mark wanted to make sure viewers could follow the thread. “So, for people who might not be familiar, how is what you’re doing different from other data scraping subnets?”
Brendan nodded. “Similar in the sense that we fetch and organize data on demand, but our edge is in scope and integrity. We scrape from sources like LinkedIn and TikTok, places most platforms can’t touch. More importantly, we enforce trusted execution environments. That means the data miners collect stays tamper-proof from source to destination.”
“That’s huge,” Mark replied. “So miners can’t just fake results or fill up the network with garbage data. Everything runs in a secure enclave. That’s a big differentiator.”
“It really is,” Brendan agreed. “It builds trust. We’re not competing on who can scrape the most data, we’re competing on who can scrape it best.”
Competition, Incentives, and Infrastructure
Mark pressed further. “Let’s talk about how your miners compete. How do you decide who wins rewards?”
“It comes down to throughput and reliability,” Brendan explained. “The miners who have the most robust infrastructure (fast, resilient systems that stay up even when platforms block scrapers) are the ones that earn more. Every data type has its own level of difficulty, so harder tasks like TikTok or financial feeds earn higher rewards.”
Mark leaned in. “So, in a sense, the miners with the beefiest rigs win.”
“Exactly,” Brendan laughed. “It’s a game of resilience. To be a good miner here, you need to be professional, fast, and ready for constant changes in the data landscape.”
From Data to Real Insights
Mark wanted to see it in action. “Can we get a quick demo?”
Brendan pulled up Gopher’s dashboard. Within seconds, it returned live results from X. “We can fetch the latest discussions about any topic (say, Bittensor or Bitcoin) and summarize them with sentiment and engagement data. We can even do this with TikTok videos, converting speech to text and analyzing top trends.”
“That’s wild,” Mark said, watching the live feed appear. “I didn’t even know you could do that with TikTok.”
Brendan nodded. “And you can. Same with LinkedIn. You can search for, say, ‘data scientists in London’ and instantly get contact insights. It’s all powered by our miner network.”
The Gopher Ecosystem and Financial AI
As the demo wrapped, Brendan revealed where Gopher was heading next. “We realized general data scraping is too broad. So we focused on finance. Our tool, Gopher Trader (GoTrader), combines real-time data scraping with predictive AI models to help users make smarter trading decisions.”
He paused, then added, “I’ve been in crypto since 2013. I built this because I needed it. I was manually pasting TradingView charts into GPT models and chatting about the market. I thought, why not automate this entire process?”
Mark grinned. “So you built the thing every trader wishes existed.”
“Pretty much,” Brendan said. “Now, we’ve got over 60,000 users and about 1,500 paying customers. The app’s predictions are hitting around a 65% success rate. It’s still early, but the traction has been incredible.”
Building on Bittensor
The conversation turned philosophical. Mark asked, “Why build Gopher as a subnet at all? You could have just built it in-house.”
“Because Bittensor gives us leverage,” Brendan said plainly. “We get 256 miners working for the network instead of paying for massive AWS infrastructure. It’s incredibly capital-efficient. Bittensor lets us align incentives across people who don’t even know each other, all working toward a single goal: reliable, verifiable data.”
Mark nodded. “It’s brilliant, really. You’re crowdsourcing infrastructure in a trustless way.”
The Token Question
Eventually, Mark raised the question every viewer was thinking. “Let’s talk about your token model. How does Gopher’s token fit into all this?”
Brendan took a measured tone. “We started with a dual-token model, and we’re still figuring out how to best align those systems. Our focus right now is on growth and building meaningful revenue. The community wants buybacks, but long-term sustainability comes first. We’re treating this like a serious company, not just a token project.”
Mark nodded approvingly. “That’s honest. No one has fully cracked the subnet token model yet. You’re figuring it out in real time.”
The Future of Subnets and Decentralized AI
As the discussion drew to a close, the two reflected on the bigger picture.
“I fell in love with Bittensor because it’s a living experiment in incentive alignment,” Brendan said. “It brings order to chaos through collaboration. Everyone’s competing, but collectively, we’re building something that works.”
Mark agreed. “It’s Darwinian: subnets evolve, the strongest survive, and everyone learns. What you’re doing with Gopher proves that.”
Brendan smiled. “We’re not building a small company. We’re building a massive entity, a decentralized AI data powerhouse.”
Closing Thoughts
As Mark wrapped up the show, he turned to the audience. “If you haven’t checked out Gopher Subnet 42 yet, you should. It’s not just another AI project, it’s a live case study in how decentralized data can power the next generation of financial intelligence.”
Brendan laughed. “Thanks, Mark. And to anyone watching, come join us. We’re just getting started.”
Be the first to comment